To monitor microbial community dynamics in a semi-industrial-scale lignocellulosic biofuel reactor system and to improve our understanding of the microbial communities involved in the MixAlco™ biomass conversion process. Reactor microbial communities were characterized at six time points over the course of an 80-day, mesophilic, semi-industrial-scale fermentation using community qPCR and 16S rRNA tag-pyrosequencing. We found the communities to be dynamic, bacterially dominated consortia capable of changing quickly in response to reactor conditions. Clostridia- and Bacteroidetes-like organisms dominated the reactor communities, but ultimately the communities established consortia containing complementary functional capacities for the degradation of lignocellulosic materials. Eighteen operational taxonomic units were found to share strong correlations with reactor acid concentration and may represent taxa integral to fermentor performance. The results of this study indicate that the emergence of complementary functional classes within the fermentor consortia may be a trait that is consistent across scales, and they suggest that there may be flexibility with respect to the specific identities of the organisms involved in the fermentor's degradation and fermentation processes. This study provides new information regarding the composition, dynamics and potential flexibility of the microbial communities associated with the MixAlco™ process and is likely to inform the improvement of this and other applications that employ mixed microbial communities. © 2010 The Authors. Journal of Applied Microbiology © 2010 The Society for Applied Microbiology.